Leveraging Big Data Analytics and Machine Learning to Predict Consumer Behavior and Enhance Marketing Performance in E-Commerce: An Integrated Predictive Framework.

Abstract

This study aims to analyze the role of big data analytics and artificial intelligence (AI) technologies in understanding consumer behavior patterns, predicting purchasing decisions, and enhancing marketing performance within the e-commerce environment. The study seeks to develop an integrated, multi-level intelligent framework that combines deep learning techniques and hybrid neural networks to extract customer behavioral patterns, predict the likelihood of purchase completion, and identify the most influential factors in consumer decision-making—thereby supporting the development of data-driven marketing strategies. The methodology involves formulating an AI-based model using neural network techniques to predict the probability of customer purchases. By identifying the most significant variables affecting purchasing decisions—utilizing Explainable Artificial Intelligence (XAI) techniques—the study compares the performance of various predictive algorithms to determine the optimal model. It further evaluates the potential for improving conversion rates, revenue, and marketing returns based on the model’s outputs. The proposed methodology centers on constructing a hybrid predictive model that integrates Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) networks, and an Attention Mechanism, alongside Explainable Artificial Intelligence (XAI) techniques. The proposed framework is based on a multi-source database that includes purchase transaction records, product browsing behavior, session duration, visit frequency, interaction history with advertising campaigns, customer demographics, and temporal data related to purchasing activities. The data undergoes preprocessing—including the cleaning of missing or anomalous values, normalization, and feature engineering—before being fed into the predictive model. The illustrative results suggest that the proposed framework successfully addresses the challenges of precisely identifying customer requirements and aspirations by accurately predicting consumer purchasing behaviorof 95.22%, a precision of 95.14%, an F1-score of 95.57%, a recall of 96.00%, and a specificity of 94.31%. This indicates that the integrated approach—which combines the extraction of spatial and temporal consumer features—offers a promising path for uncovering consumer behavior and enhancing purchasing processes. Furthermore, by incorporating the proposed framework into decision-support and intervention modules, it can serve as an effective means to improve marketing operations.